Machine learning is one of the important components of the technology traditionally referred to as artificial intelligence. Machine learning uses statistical methods to learn autonomously, that is, without direct instruction from designers and without the use of programming.
The basis for machine learning is algorithms, which are a complex set of mathematical models and rules based on which AI analyses information, draws conclusions, makes decisions and makes predictions.
Machine learning methods
- Supervised learning is a method of using algorithms on labelled training data. Artificial intelligence is trained on a trained database and then extrapolates the lessons learnt to predict the situation in new scenarios.
- Unsupervised learning – the use of algorithms on an untrained database. Artificial intelligence analyses the information, independently identifying structures, relationships and patterns.
- Reinforcement learning – algorithms learn in a certain environment on the basis of their own actions by trial and error. In this case, the principle of reward maximisation is relevant, when the machine strives to get the best result based on the experience gained.